Imagine if in 2006, just when U.S. housing prices had reached their peak, the country’s largest banks had been required to stress test their loan and securities portfolios to determine what would happen in the seemingly far-fetched scenario that home values plummeted.
There’s no way of knowing if things over the past five years would have been all that different—if it would have lessened the impact of the devastating decline in housing values on bank balance sheets and the economy as a whole. But odds are, it wouldn’t have hurt, especially if lenders had been forced to incorporate the results of those tests into their long-term plans by clamping down on lending risk and setting aside extra capital to protect themselves from the effects of a real estate price collapse.
Perhaps we wouldn’t have experienced a global liquidity crisis and the disruptive failures or forced sales of several prominent U.S. financial institutions. Perhaps the economy wouldn’t have suffered quite the same ill effects from the housing meltdown. Perhaps banks wouldn’t have needed government bailouts or become punch lines for the jokes of late-night comedians.
Preventing the next financial crisis is the rationale behind the latest spate of congressional and regulatory requirements that the country’s largest institutions undergo annual stress tests. Before long, experts say, all banks will be expected to regularly “shock” their balance sheets and income statements to see what would happen if things don’t go as well as planned.
Not surprisingly, the requirement is eliciting howls of protest from bankers of all shapes and sizes. To critics, it’s the latest in a series of costly—and unnecessary—government mandates seemingly meant to make banking as time-consuming and unprofitable as possible.
But one banker who’s not at all disturbed is Robert Jones, the chief executive officer at $8.9-billion asset Old National Bancorp in Evansville, Indiana. Indeed, Jones considers the practice of “shocking” loan portfolios and income statements for such “what-if” scenarios as interest-rate spikes, falling housing prices or even changes in regulatory policy to be a key—and necessary—part of his strategic arsenal. If the government wants to require the testing to get a better handle on risks, well, there have been worse ideas.
“Quite frankly, I think stress testing is just good management—the definition of good governance,” Jones says. “It allows the board to get comfort that we’re moving in the right direction in multiple areas of the bank, and that we have the capital and earnings to provide consistency to our shareholders. We find it to be extremely valuable.”
Jones has the necessary attitude to take on what’s coming next. Three years after the Supervisory Capital Assessment Program (SCAP) mandated stress testing for the nation’s 19 largest financial institutions, the process is going mainstream.
Those same 19 firms are now undergoing a third round of Federal Reserve-administered testing under the Comprehensive Capital Analysis and Review (CCAR) process.
The process and purpose of the tests have evolved regularly since SCAP arrived on the scene in 2009. SCAP employed government-administered stress tests in an effort to restore public confidence in large banks by ensuring that they had enough capital to hold up in a stressed economic environment. The results were made public, and institutions found lacking were required to raise additional capital.
The 2011 tests under CCAR were self-administered by the same 19 institutions using a Fed-devised scenario, and were meant to assess their capital planning processes.
“If all we ever did was run supervisory tests in which we instruct banks in detail how to perform the test, we would be in the position of a parent who shows his child how to solve each problem in her homework, and never discovers whether the child can do the work on her own or not,” explained William Dudley, CEO of the New York Fed, in a June 2011 speech to a group of global financial leaders in Bern, Switzerland. “We want banks to have good capital-planning processes and make intelligent decisions with respect to capital.”
The 2012 round expands the number of institutions undergoing review to those with more than $50 billion in assets, adding the likes of M&T Bank Corp., Zions Bancorporation and Comerica Inc. to a list that already includes the industry’s titans.
One of the main goals this time around is to ensure that holding companies can maintain a Tier 1 capital ratio of 5 percent under a scenario where the economy experiences a deep recession, with a 13 percent unemployment rate and a GDP decline of 8 percent, while markets abroad also contract. If regulators don’t like what they see, they will have the power to nix stock dividend or repurchase plans.
“The key difference is that the capital distribution decision now hangs on the stress test, as opposed to it just being a best practice,” says Karen Shaw Petrou, managing partner of Federal Financial Analytics, a Washington, D.C., consulting firm. “Instead of the Fed needing to prove why the bank shouldn’t make a capital distribution, the bank will have to prove why it should.”
The six largest institutions, including JPMorgan Chase & Co., Bank of America Corp., Wells Fargo & Co. and Citigroup, also must undergo additional tests to gauge how they might survive a “hypothetical global market shock” rooted in Europe’s recent sovereign debt troubles.
In a parallel vein, the Dodd-Frank Act requires all banks with more than $50 billion in assets to undergo Fed-run stress tests. Those institutions also are expected to conduct their own semi-annual stress tests and report the results to their primary regulators. While the mechanics of the Dodd-Frank tests have yet to be finalized, “I expect that the SCAP and CCAR exercises will serve as models,” Dudley said.
Confused? Get used to it. Like it or not, stress-testing looks like it’s here to stay—and not only for the big banks. The next tier of banks—those with assets between $10 billion and $50 billion–won’t have the Fed running any stress tests, but Dodd-Frank does require them to perform tests of their own. That notion has been seconded by the three federal regulatory agencies, which last June issued proposed guidance similarly calling for such do-it-yourself tests for the $10 billion-and-over crowd.
Banks below the $10-billion asset threshold are technically exempt from the requirement, and they’d like to keep it that way. “We think it would be overkill for community banks,” says Chris Cole, chief regulatory counsel for the Independent Community Bankers of America.
Even Cole concedes that he’s likely fighting a losing battle. Indeed, banks in the $100-million asset range already are being told to gear up for testing. “The examiner-in-charge, on the way out the door, will say, ‘By the way, we think you should be doing this,’” says Peter Cherpack, a senior vice president and director of credit risk process and technology for Ardmore Banking Advisors, an Ardmore, Pennsylvania, provider of stress testing systems.
“The expectations for larger banks almost always spill over to the rest of the system. In this particular exercise the evolution has been quite rapid,” explains Sabeth Siddique, a Washington, D.C.-based director of Deloitte & Touche LLP’s governance, regulatory and risk-strategies practice.
While the exact extent is still unknown, “the examiners are going to be looking for stress testing in all banks,” adds Siddique, a former assistant of banking supervision and regulation at the Federal Reserve. “I know that for a fact.”
Bankers don’t like to talk much about the government’s new stress-testing regime. Of nearly a dozen larger bank representatives contacted for this story, none was willing to be quoted on the subject.
(Representatives from the Fed, the Office of the Comptroller of the Currency and the Federal Deposit Insurance Corp. also declined or did not respond to interview requests.)
Privately, big-bank executives complain that it’s too harsh and arduous, and requires too much time and money to perform in the proper manner. Tom Brown, founder of Second Curve Capital, a New York-based hedge fund that specializes in financial services, calls the idea of annual stress tests “an incredibly stupid, truly nonsensical practice.”
A former bank analyst who also publishes the Bankstocks.com website, Brown notes on his blog that banks are subject to minimum capital requirements precisely to get them through periods of stress. If regulators are worried that banks might not hold up under difficult conditions, “the response should not be to layer in a new round of annual stress tests,” he says. Rather, Washington should “dial up the minimum amount of capital cushion that banks have to hold.”
Siddique, one of the architects of the Fed’s first SCAP tests, says that’s a nice idea in theory. The reality, he says, is that different banks hold wildly different types of assets; getting the capital levels right demands more individualized attention—something that can be achieved through stress testing. “The riskiness of banks’ portfolios varies quite a bit,” he says.
Banks have long made regular use of stress tests as part of their broader enterprise risk-management approach, and to gauge liquidity risks and plot their capital and operational strategies. Those that have done them have typically shared the results with their regulators, although with hindsight many of those efforts have been criticized for focusing more on threats to profitability than survival.
The government requirement both changes the emphasis and ratchets up the pressure.
Under the CCAR big-bank stress-testing regime, if a bank can’t demonstrate the ability to survive really trying conditions, then regulators will rule out dividends, and the bank will likely be required to raise additional capital and/or change some of the ways it does business.
If, on the other hand, a financial institution shows it has the capital strength to ride out adverse conditions, then everyone—regulators, investors and customers—will be more comfortable.
Preserving system stability by minimizing the risk of failures is the core objective of required stress testing. At the institution level, capital adequacy and liquidity are the names of the game.
Banks that run into capital problems often see funding evaporate. It’s a lack of liquidity that ultimately takes an institution down.
“Most institutions don’t fail directly from credit or earnings, they fail because of funding,” explains Jeff Curry, a managing director for FTI Consulting, a West Palm Beach, Florida, firm that has helped several big banks set up stress tests.
“The regulators want to see banks become more proactive in how they manage risk,” Curry adds. “Stress testing is one of the best tools available for getting that kind of forward-looking view.”
Stress testing is conceptually easy to understand. The basic idea is to try to envision how things might go wrong, and then figure out what it would do to the bank’s balance sheet and income statement—and what might be done to prepare for the worst.
Curry offers a simple example of liquidity stress testing, where three different bank-specific scenarios are contemplated. In the first, bad earnings, harsh regulatory action or other negative publicity spark customer concern, and 7 percent of deposits flee. A second, more severe scenario leads to a 15 percent loss in deposits, including some local government accounts.
A third “worst-case” scenario sees deposits plunge by 25 percent. “You have commercial depositors withdrawing DDAs. Wholesale funding gets shut down and the Home Loan Bank won’t give any more advances because of collateral levels,” Curry says.
In each of those cases, the stress test would reveal things such as the bank’s loan-to-deposit ratio, how much on-balance sheet liquidity it would have and how much it could garner selling the securities in its portfolio.
If the picture looks glum—and even though the scenario might seem far-fetched—the board might feel heat from the regulators to shore things up, say, by securing an additional borrowing line, raising more capital, divesting a business or changing underwriting standards on some loans to lessen the risks.
While the Fed has laid out specific stress-test scenarios for big banks, there’s no “right” way to get the task done, experts say. Every institution has its own loan compositions and exposures, and thus its own unique risks.
The scenarios employed can be broad in nature: What would happen to the bank if the unemployment rate spiked to 12 percent? What if housing prices fell another 20 percent? Or they can be more specific. What if the biggest employer in town shut down? An agricultural bank might want to figure out what would happen to its loan portfolio if commodity prices fell by 20 percent.
In all cases, however, the end questions should be the same: Does the bank have enough capital to weather the hypothetical storm? Is there anything the board and management can do now to proactively minimize the potential damage?
Stress testing “gives you an idea of what could happen, so you can plan for contingencies, as opposed to blundering into a situation and having to figure things out on the fly,” Cherpack explains.
The proposed guidance, which is expected to become official in early 2012 with few changes, lays out a set of principles and methods for proper testing. But it also is somewhat short on specifics, and “is not intended to provide detailed instructions,” which experts say is a good thing.
The principles are straightforward, and important for all board members to understand:
Tailor what’s tested to the institution.
Employ “multiple stress testing activities and approaches, and ensure that each is conceptually sound.”
Make the tests “forward looking and flexible.”
Make the results clear and actionable, and use them to “inform decision-making.”
Banks should “strongly consider” using a combination of four types of testing—“scenario analysis,” “sensitivity analysis,” enterprise-wide testing and “reverse testing”—to gain that understanding, the guidance adds. Reverse testing involves working backward from a failure position to determine what it would take to get there.
The numbers don’t need to be perfect. By definition, the tests include a variety of what-ifs, guesses and approximations, Cherpack advises. “When you run with numbers like that, the result is purely directional,” he says. “It’s not the answer.”
It sounds intimidating. With so much work—and so much riding on the results—there’s plenty of stress in bank boardrooms over the government’s stress-testing requirement.
Regulators estimate that it would take a typical bank about 260 hours to collect the information needed for stress tests under the proposed guidance. A comment letter on the proposed guidance submitted by four trade groups, including the Financial Services Roundtable and American Bankers Association, counters that the tally would be something closer to “several thousand hours” for an average bank.
For the biggest banks, “it’s a great deal of work” to pull such tests together, Petrou says. Some banks have created independent “model validation teams,” for instance, to sign off on the models crafted by the risk and finance people; others use big consulting firms for the task.
Even for smaller banks, the amount of work can be substantial. Since stress testing is driven by data, many banks must update appraisal values on collateral properties and other key financial information–a process that doesn’t come cheap. “If your appraisals are four years old, and commercial real estate prices have dropped 28 percent during that time, you need to get those updated,” Cherpack says.
As for the calculations themselves, a small community bank might be able to do them on a spreadsheet. Most others likely will need something more sophisticated. Many banks have begun turning to outside technology vendors and consultants to make sure the job is being done right.
For a community bank, the tab for one of Ardmore’s solutions starts at $5,000, Cherpack says. For larger banks with complicated operations, solutions from big vendors like SunGard, Oracle Corp. or International Business Machines Corp. can reach into the millions.
Director workloads are almost certain to increase in response to yet another regulatory mandate. Ron Glancz, chairman of the financial services group at Venable LLP, a Washington, D.C., law firm, has heard the grousing: “Here’s one more burden that we’ve got to deal with. We need to focus on the business and making money, not another big round of reporting requirements,” he says. “It feels like piling on.”
That may be, but it’s also part of the job. “If you’re not comfortable reviewing the results of a stress test—if it’s too hard or too complex—then don’t be a bank director,” Petrou says. “Some of the requirements boards are asked to supervise aren’t franchise-critical. This one is.”
For banks over $10 billion, the risk committee of the board is generally charged with oversight of the stress testing process. Smaller institutions that aren’t required to have a risk committee might assign the task to the audit or asset/liability committee. But all directors will be expected to participate in the process.
“They don’t need to get into the weeds, but they should know the tail risks [and] what the downside exposures of the bank are,” Curry says. “And they should have a strong dialogue with management. This shouldn’t be a one-way communication.”
The proposed guidance says the board must ensure that principles are adhered to, and that the findings are integrated into decision-making at the business line, capital-allocation and asset/liability levels. Directors should come out of the exercise confident that the company’s risk profile matches up with “the board’s chosen risk appetite.”
As with other key governance areas, directors will be expected to ask plenty of questions and be ready to incorporate test results into business- and capital-planning efforts. What assumptions are being used in the process? Have real estate collateral values been updated to present an accurate picture of the portfolio?
If board members don’t understand something, they should seek plain-English explanations on both the inputs and results. “The whole process requires that the board fully understands the strengths and weaknesses of the capital-adequacy process,” Siddique says. If, for instance, directors suspect that certain models might be underestimating losses, then they “might want to add a [capital] buffer to account for that.”
Stress testing done right is more than just a quantitative exercise, he adds. It often leads to subtle changes in a bank’s culture, because information has to be shared and utilized in different ways.
“Everyone—the quant people, the risk people, treasury, the business-line people—needs to work together to create one consistent story,” Siddique says. “This exercise should, indirectly, change the way banks are structured internally, and force people in the bank to work more closely together.”
That carrot is reinforced with a stick. Directors of banks that don’t perform stress tests, or do them in a “haphazard” manner, could be subject to enforcement actions or civil money penalties, Glancz warns. “If the bank fails, it would be a basis for the FDIC receiver to say directors had breached their fiduciary duties,” he adds.
The good news in all this is that many banks are already broken in on the process, at least to some extent, and have found that it can be beneficial.
Recipients of government capital from the Troubled Asset Relief Program (TARP) have been required to stress their balance sheets before being allowed to buy their way out of the program, for instance. And regulatory guidance issued in 2006 by the federal banking agencies still requires stress tests for banks with commercial real estate exposures above 300 percent of capital or construction exposures above 100 percent of capital.
Many banks conduct regular stress tests as a best practice, using the information gleaned from the exercise to look ahead and prepare strategic responses to changes in the marketplace.
Old National’s first brush with serious stress testing actually came at the board’s behest. The company took $100 million in TARP money in December 2008, and three months later was among the first to buy its way out. Stress tests weren’t required yet for TARP exits, but “the board asked for it” anyway, says Jones, the bank’s CEO. “Directors wanted to be sure that we wouldn’t need that capital in a worst-case scenario.”
With help from investment bank Sandler O’Neill + Partners in New York, it fashioned a “draconian” scenario of a global economic slowdown, “and we were able to show the board that we had more than sufficient capital,” Jones says.
Nowadays, Old National stress tests every quarter. Jones’ team plugs in a variety of economic scenarios, including both best guesses and worst cases, to see how credit quality, liquidity and earnings would be affected. It then incorporates those findings into key strategic and capital-planning decisions, including acquisitions.
The practice has paid important dividends. Two years ago, Old National used a stress test to calculate how much fee income would be lost if a rumored clampdown on account services fees took effect. “As we looked out, we just couldn’t find a way to make the product profitable the way it was designed,” Jones recalls.
When an amendment to Regulation E, which imposed new restrictions on bad check fees, became law, the company was prepared with a plan to charge for checking accounts. It introduced the changes quickly, and unlike some large banks, the modifications stuck. The company predicted attrition rates in its model, and those projections have held up, too. “Our checking balances continue to grow, and our profitability has grown as well,” Jones says.
“We made the decision based on that test,” Jones adds. “In the old days, before stress testing, we couldn’t have done that. We would have just held our fingers to the wind, hoping we made the right decision.”